Portal de Periódicos da CAPES

A Place and a Grid in the Sun

2014; Cell Press; Volume: 159; Issue: 6 Linguagem: Inglês

10.1016/j.cell.2014.11.033

1097-4172

Eric R. Kandel,

Sleep and Wakefulness Research

The 2014 Nobel Prize in Physiology or Medicine, awarded to John O’Keefe, May-Britt Moser, and Edvard I. Moser, recognizes the first deep-brain insights into a cognitive function. Their insights established a new view for how the brain represents spatial location. The 2014 Nobel Prize in Physiology or Medicine, awarded to John O’Keefe, May-Britt Moser, and Edvard I. Moser, recognizes the first deep-brain insights into a cognitive function. Their insights established a new view for how the brain represents spatial location. This year the Nobel Prize Committee honored the discoveries, by John O’Keefe and by Edvard and May-Britt Moser, respectively, of the neural encoding of spatial location through place cells and grid cells. These groundbreaking contributions to cognitive neuroscience provided our first understanding of the role of the hippocampus and the entorhinal cortex in the brain’s representation of space, which is likely to be critical for an animal’s ability to form associative memories between a particular location in its environment and a particular sensory context. Their studies have also given us our first insights into a cognitive representation in higher brain regions that is concerned not simply with primary sensory or motor representation, but with a complex and abstract combination of modalities. Classic studies by Brenda Milner and William Scoville in 1957 established the importance of the hippocampus for encoding declarative memory, the recall of information about people, places, objects, and events. Despite this advance, no one knew anything about the sensory signals that activated hippocampal neurons; in particular, no one thought the hippocampus might be important in the representation of space. As a result, it came as a great surprise when, in 1971, John O’Keefe and his student Jonathan Dostrovsky discovered that the pyramidal cells of the hippocampus encode not a single sensory modality—not touch, vision, taste, or smell—but something abstract, a representation, or cognitive map, of space (O’Keefe and Dostrovsky, 1971O’Keefe J. Dostrovsky J. Brain Res. 1971; 34: 171-175Crossref PubMed Scopus (3822) Google Scholar, O’Keefe and Nadel, 1978O’Keefe J. Nadel L. The Hippocampus as a Cognitive Map. Oxford University Press, Oxford1978Google Scholar). O’Keefe and Dostrovsky discovered that various hippocampal neurons in the CA1 region of a rat’s brain fired briefly when the animal assumed different positions in space. O’Keefe called these neurons “place cells” and the location that triggers the firing of each cell its “place field.” By recording from many place cells at the same time, he further noticed that the firing of a given cell or set of cells forms a map indicating where in space the animal is located (O’Keefe, 1976O’Keefe J. Exp. Neurol. 1976; 51: 78-109Crossref PubMed Scopus (1160) Google Scholar, O’Keefe and Conway, 1978O’Keefe J. Conway D.H. Exp. Brain Res. 1978; 31: 573-590Crossref PubMed Scopus (678) Google Scholar). When O’Keefe moved the rat to another space, the animal would form a new map using some of the same place cells as well as some other place cells. When he brought the rat back to the original space, the initial cells would fire again, reforming the initial map. O’Keefe went on to demonstrate that the orientation of this map relative to an environment can be changed from trial to trial, and that this orientation is “remembered” even without spatial cues (O’Keefe and Speakman, 1987O’Keefe J. Speakman A. Exp. Brain Res. 1987; 68: 1-27Crossref PubMed Scopus (556) Google Scholar). O’Keefe’s second major insight was that the organization of the spatial map in the hippocampus is radically different from that of the sensory maps for touch and vision in the cortex. The maps for touch and vision are typically organized topographically, as Wade Marshall, Vernon Mountcastle, David Hubel, and Torsten Wiesel had discovered earlier; that is, neighboring cells in the cortex convey information about neighboring areas of the sensory periphery. The brain’s spatial map is not organized topographically; thus, neighboring cells in the hippocampus do not convey information about neighboring positions in space. In fact, the spatial map is random with no topographic relation between neighboring cells and the part of external space they represent (O’Keefe and Nadel, 1978O’Keefe J. Nadel L. The Hippocampus as a Cognitive Map. Oxford University Press, Oxford1978Google Scholar). This discovery, too, was totally unexpected and represents an entirely new view of how information is organized in the mammalian brain. It is generally considered the most important new finding in the sensory physiology of the mammalian brain since the Hubel-Wiesel finding of linear receptive fields in the primary visual cortex. O’Keefe’s third major contribution was theta phase precession—the finding that neurons in the hippocampus encode location in space through both a rate code (firing frequency) and a temporal code (when the neurons fire). The hippocampus generates a coordinated pattern of activity as an animal explores its environment. This activity is produced by a network of neurons and gives rise to an extracellular electrical signal measured in an EEG called the theta rhythm, or oscillation, which ranges in frequency from 4 to 12 Hz. In 1993, O’Keefe and Michael Reece found that as an animal moves through a given neuron’s place field, the timing of spike firing shifts progressively to earlier phases of the theta oscillation (O’Keefe and Recce, 1993O’Keefe J. Recce M.L. Hippocampus. 1993; 3: 317-330Crossref PubMed Scopus (1783) Google Scholar). Because the place fields of different cells overlap, each place cell will fire at a different phase of the theta oscillation, enabling the brain to determine the animal’s position with good precision (Huxter et al., 2003Huxter J. Burgess N. O’Keefe J. Nature. 2003; 425: 828-832Crossref PubMed Scopus (444) Google Scholar). This phenomenon of phase precession provides one of the few well-characterized examples in which the brain uses “temporal coding” that is not directly associated with the temporal properties of the stimulus. When May-Britt and Edvard Moser entered the field, they focused on analyzing how the neural circuitry of the hippocampus relates to spatial behavior. The hippocampus receives two types of connections from the entorhinal cortex. One is via the direct pathway that carries information from the entorhinal cortex to area CA1 of the hippocampus. The other is via the indirect pathway from the entorhinal cortex via the dentate gyrus and the CA3 region of the hippocampus. They discovered that the brain’s representation of space develops independently in areas CA3 and CA1 of the hippocampus. They then went on to show that the direct pathway from the entorhinal cortex to area CA1 is sufficient to maintain a spatial map: for the recognition of space; the indirect pathway is not required (Brun et al., 2002Brun V.H. Otnass M.K. Molden S. Steffenach H.A. Witter M.P. Moser M.B. Moser E.I. Science. 2002; 296: 2243-2246Crossref PubMed Scopus (486) Google Scholar). Having demonstrated that the spatial map is not intrinsic to the hippocampus and that the direct pathway is important, the Mosers then asked: how is the spatial map formed from this direct input? Is there a code for space in the entorhinal cortex, the major route of excitatory input to the hippocampus? In 2004 the Mosers discovered that, indeed, certain neurons in the medial part of the entorhinal cortex do encode an animal’s location in space (Fyhn et al., 2004Fyhn M. Molden S. Witter M.P. Moser E.I. Moser M.-B. Science. 2004; 305: 1258-1264Crossref PubMed Scopus (902) Google Scholar). However, unlike the place cells in the hippocampus, each of which encodes a unique place field, neurons in the entorhinal cortex represent space in a grid-like firing pattern (Hafting et al., 2005Hafting T. Fyhn M. Molden S. Moser M.-B. Moser E.I. Nature. 2005; 436: 801-806Crossref PubMed Scopus (2437) Google Scholar). Each cell fires not in a single location in space, but in multiple, evenly spaced locations. These locations form a periodic triangular, grid-like array, with clear regions of silence between the vertices of the triangles. As the Mosers expanded the size of the environment the rat explored, they found that the grid-like firing field of each individual entorhinal neuron repeats itself across the entire space. Each cell imposes its pattern of firing on every environment the animal encounters. Thus, while place cells in the CA1 region provide a local description of space, grid cells provide a global description (Hafting et al., 2005Hafting T. Fyhn M. Molden S. Moser M.-B. Moser E.I. Nature. 2005; 436: 801-806Crossref PubMed Scopus (2437) Google Scholar, Fyhn et al., 2007Fyhn M. Hafting T. Treves A. Moser M.-B. Moser E.I. Nature. 2007; 446: 190-194Crossref PubMed Scopus (471) Google Scholar). The finding of grid cells helped pinpoint the entorhinal cortex as a key hub in the brain network that enables us to find and remember our way. The Mosers next searched for additional types of cells in the entorhinal cortex that might code for space. They found two: head direction cells (first described by Jim Ranck in the subiculum) that respond to the direction of the animal’s head with respect to the environment and border cells that respond to the presence of a border or edge in the environment (Sargolini et al., 2006Sargolini F. Fyhn M. Hafting T. McNaughton B.L. Witter M.P. Moser M.-B. Moser E.I. Science. 2006; 312: 758-762Crossref PubMed Scopus (930) Google Scholar, Solstad et al., 2008Solstad T. Boccara C.N. Kropff E. Moser M.-B. Moser E.I. Science. 2008; 322: 1865-1868Crossref PubMed Scopus (697) Google Scholar). These two cell types generate a continuously updated representation of the position of the animal in space that can be used in any environment, irrespective of shape and landmarks. These findings raised the question: Are the inputs from the grid cells in the entorhinal cortex the only ones that contribute to the formation of hippocampal place cells? The Mosers next went on to provide evidence that place cells are most likely generated by the convergence of signals from all three entorhinal cell types. The first evidence for this convergence emerged when the Mosers and John O’Keefe independently examined the causal connection between grid cells and place cells during development. They argued that if place cells are generated exclusively from grid cells, then grid cells should be active before, or at least simultaneously with, the emergence of place cell activity. However, recordings from rat pups revealed that this is not the case (Langston et al., 2010Langston R.F. Ainge J.A. Couey J.J. Canto C.B. Bjerknes T.L. Witter M.P. Moser E.I. Moser M.-B. Science. 2010; 328: 1576-1580Crossref PubMed Scopus (416) Google Scholar). At 2 weeks of age, when rats first leave the nest, a large proportion of hippocampal cells already have place fields and therefore function as place cells, whereas grid cells exhibit only weak periodic fields. This suggests either that the weak spatial periodicity of the grid cells is sufficient to generate well-defined place fields in the hippocampus, or, more likely, that place cells are formed from the combined activity of grid cells and the two other types of entorhinal cells, particularly border cells, which exhibit some properties of adult cells at 2 weeks. These results suggested that individual place cells receive important inputs from both grid cells and border cells, with grid cells providing information about distance based on the animal’s motion and border cells providing the animal’s position in relation to geometric boundaries. The strongest input seems to be from grid cells, which are several times more abundant than border cells, but under most circumstances the two classes of input are coherent and redundant: if one is absent, the other is sufficient to generate localized firing of place cells in the hippocampus. Finally, since the hippocampus is known to send a reciprocal output to the entorhinal cortex, the Mosers examined how inactivation of the hippocampus influences entorhinal activity (Bonnevie et al., 2013Bonnevie T. Dunn B. Fyhn M. Hafting T. Derdikman D. Kubie J.L. Roudi Y. Moser E.I. Moser M.B. Nat. Neurosci. 2013; 16: 309-317Crossref PubMed Scopus (237) Google Scholar). They found that silencing the hippocampus leads to a loss of grid cell periodicity and transforms grid cells into head direction cells. Thus, the encoding of space is not generated by a hierarchical, linear array of synaptic connections from the entorhinal cortex to the hippocampus, but rather depends on a recurrent loop of interconnections from the entorhinal cortex to the hippocampus and back to the entorhinal cortex. The discoveries of grid cells by the Mosers and place cells by O’Keefe provide us with a new view of how the brain represents position in space. This spatial map, as I have emphasized, is a true cognitive function. There is no single modality, no single sensory organ, that determines location. Space is computed through a number of different sensory modalities—vestibular input, touch, vision, and smell—and it represents the coordinated totality of that input. In this context, grid cells are particularly interesting because they are constructed globally from these sensory inputs. John O’Keefe received his Ph.D. from McGill University in 1967, when cognitive neuroscience was just emerging. The chairman of the psychology department at that time was Donald Hebb, who had studied with Karl Lashley and was exploring the problems of spatial orientation and spatial learning. Hebb went on to write “The Organization of Behavior: A Neuropsychological Theory,” a classic text in which he argued that all behavior must be studied through the brain. Hebb, in turn, influenced the psychologist Brenda Milner, who, with the neurosurgeon Wilder Penfield, examined the hippocampus and its role in memory storage. As a postdoctoral fellow, O’Keefe worked with Patrick Wall at University College London. There, together with Dostrovsky, O’Keefe discovered place cells by analyzing the various factors that influence the firing properties of hippocampal neurons. Together with Lynn Nadel, O’Keefe appreciated the historical importance of Edward Tolman’s ideas. One of the earliest cognitive psychologists, Tolman argued in 1948 that rats running in a maze acquire “something like a field map of the environment” (Tolman, 1948Tolman E.C. Psychol. Rev. 1948; 55: 189-208Crossref PubMed Scopus (3681) Google Scholar). Moreover, Tolman held that stimuli influence behavior not through stimulus-response connections, as Skinner and the behaviorists argued, but through the mediation of a cognitive map. Tolman thought of cognitive maps as broad, comprehensive maps or narrow strip maps confined to knowledge of specific groups. O’Keefe’s finding of a map was interesting enough in its own right, but of even greater interest was the fact that it is organized through a random population of cells. The idea of random organization of neural populations is now emerging as a broad theme in studies of higher regions of the cortex, for example the piriform cortex and the sense of smell (Stettler and Axel, 2009Stettler D.D. Axel R. Neuron. 2009; 63: 854-864Abstract Full Text Full Text PDF PubMed Scopus (380) Google Scholar). May-Britt and Edvard Moser were students of Per Andersen, an early pioneer in hippocampal electrophysiology, at the University of Oslo. There, the Mosers met and married, and began their long-lasting successful personal and scientific collaboration. Indeed, with this award the Mosers became one of a select group of married couples, including Marie and Pierre Curie, to receive the Nobel Prize. With Per Andersen, the Mosers studied mechanisms of memory formation in the hippocampus in freely moving animals. In 1995–1996, they worked as postdoctoral fellows with Richard Morris, investigating the role of long-term potentiation in hippocampal memory. Then in 1996, they spent a brief but productive three month period with John O’Keefe, learning tetrode recording in the hippocampus. This experience started them on a quest that resulted in the remarkable discovery of grid cells, which linked the early classical work on the hippocampus to later studies of the hippocampus’s input region in the entorhinal cortex. Thus, the Mosers are distinctive not only because of their discovery of entorhinal grid cells, border cells, and head direction cells—and the role these cells play in the formation of place cells—but also because of their important contribution to the physiology of the hippocampus (Leutgeb et al., 2007Leutgeb J.K. Leutgeb S. Moser M.B. Moser E.I. Science. 2007; 315: 961-966Crossref PubMed Scopus (1099) Google Scholar). Such intellectual continuity represents science at its best, and this level of excellence is made even better by the fact that these three remarkable scientists are also extraordinarily fine human beings. O’Keefe is currently Director of the Sainsbury Wellcome Center for Neural Circuits and Behavior at University College London. May-Britt Moser is Director of the Center for Neural Computation in Trondheim. Edvard I. Moser is Director of the Kavli Institute for Systems Neuroscience in Trondheim.Edvard I. Moser. Image courtesy of Geir Mogen/NTNU.View Large Image Figure ViewerDownload Hi-res image Download (PPT)